Slashdot Mirror
Linux Kernel 4.9 Officially Released (kernel.org)
"As expected, today, December 11, 2016, Linus Torvalds unleashed the final release of the highly anticipated Linux 4.9 kernel," reports Softpedia. prisoninmate shares their article:
Linux kernel 4.9 entered development in mid-October, on the 15th, when Linus Torvalds decided to cut the merge window short by a day just to keep people on their toes, but also to prevent them from sending last-minute pull requests that might cause issues like it happened with the release of Linux kernel 4.8, which landed just two weeks before first RC of Linux 4.9 hit the streets... There are many great new features implemented in Linux kernel 4.9, but by far the most exciting one is the experimental support for older AMD Radeon graphics cards from the Southern Islands/GCN 1.0 family, which was injected to the open-source AMDGPU graphics driver...
There are also various interesting improvements for modern AMD Radeon GPUs, such as virtual display support and better reset support, both of which are implemented in the AMDGPU driver. For Intel GPU users, there's DMA-BUF implicit fencing, and some Intel Atom processors got a P-State performance boost. Intel Skylake improvements are also present in Linux kernel 4.9.
There's also dynamic thread-tracing, according to Linux Today. (And hopefully they fixed the "buggy crap" that made it into Linux 4.8.) LWN.net calls this "by far the busiest cycle in the history of the kernel project."
There are also various interesting improvements for modern AMD Radeon GPUs, such as virtual display support and better reset support, both of which are implemented in the AMDGPU driver. For Intel GPU users, there's DMA-BUF implicit fencing, and some Intel Atom processors got a P-State performance boost. Intel Skylake improvements are also present in Linux kernel 4.9.
There's also dynamic thread-tracing, according to Linux Today. (And hopefully they fixed the "buggy crap" that made it into Linux 4.8.) LWN.net calls this "by far the busiest cycle in the history of the kernel project."
The kernel is pretty damn useless if the userland is inaccessible because the init system broke unexpectedly and the system didn't boot properly.
An OS needs a robust kernel, a robust init system and a robust userland. If even just one of those isn't working right, the entire OS is useless.
You don't know what a kernel is, do you?
The kernel is only a very small part of systemd at this point. Nothing to get too worked up about now. Wait until kerneld and kernelctl get to version 1.0 ...
It must have been something you assimilated. . . .
I think someone needs to "make the init system great, again"
(sorry, lol)
--
"It is now safe to switch off your computer."
Hahahahahahah. Try opening a terminal. Now try opening a tab within the terminal.
Hahhahahahha.
The kernel is pretty damn useless if the userland is inaccessible because the init system broke unexpectedly and the system didn't boot properly.
Duh, "apt purge systemd" and you can enjoy a reliable init. Just like the solution for most sound problems is "apt purge pulseaudio". Or, closing a link-local security hole by "apt purge avahi-daemon". I think you get the pattern.
The creatures outside looked from Alt-Right to Antifa; but already it was impossible to say which was which.
The damage is already done. The paradigm has already shifted from, "I'm a Unix guy so I understand what the machine is doing" to, "Nobody understands what the machine is doing". Any time you can take Poettering software out of the loop, your system becomes a lot more sane and understandable. But, it's becoming increasingly more difficult to do that and, with the deep pockets of RedHat, they will eventually own the new dystopian Linux userspace. RedHat does a lot of good in the Linux world but, the badness they have unleashed is almost unforgivable.
Or just go back to sysvinit. It worked perfectly fine for a long time, and there was no good reason to switch to systemd.
I don't have much firsthand knowledge here but I suspect this is why systemd has seen such widespread adoption despite its warts. If you don't care about dependency management via declarative syntax that's fine, but distro builders probably appreciate it. If you view its ability to babysit processes as un-UNIXy and a slippery slope leading towards a more Windows-like state of affairs, heck I'd largely agree with you there... but once again, this is something distro builders and many other developers are going to quickly become addicted to. If you think virtualization/containerization features are useless, you're living in the stone age.
You greybeards could have halted this thing in its tracks early on by throwing your weight behind an alternative like OpenRC, but instead the majority appeared to adopt this "there is absolutely no reason to care about any systemd feature whatsoever" attitude... so now we're stuck with the bad solution dominating the ecosystem unless and until enough people can throw their weight behind good solutions and play catchup... and yet people like you are still stuck on the "no solution required, damnit!" point of view.
The psychological parallels between the systemd debacle and the Trump debacle are surprisingly deep. Both are cases where the existence of dire flaws in a proposed solution caused the judgment centers in brains of most detractors (or at least the most of the vocal detractors) to short-circuit.